Not quite. Your knight belongs on the f3 square, not the pawn

As per general opening traditions, a knight on f3 would take partial control on d4 and exerts pressure on black e-pawn. Another issue of this move pattern is the white LSB is blocked by his own pawns. Instead of 4.d3, 4.Nge2 and 5.d4 would make a 'basic recovery'.

I think he is only pinning the knight, not really intending to take it, because then he would lose a bishop for a knight.

Usually one does not want to concede a pin and the bishop pair so early...

Better, in my view, were 6.Bh4 or even 6.Be3. Even though the latter choice constitutes a bit of a loss of time (5.Be3 got there a move faster), that you have induced ...h6 might be considered worth it. After taking the Knight, White might start having problems with the dark squares.

I may be wrong, but here again, trading a bishop for a knight is not the best.

I actually disagree with this capture. White's dark squares are awfully weak and without the Dark-Squared Bishop to defend it, cannot be defended easily. I would've kept this bishop even if it meant 7...Bc5 8. Nd5 Qd8

I second it, black has just given away a precious bishop merely for an exchange. If black's last moves were something like Bc5, d6 and Be6, it might be a prelude a fierce middle game for black possessing superior bishops than opponent.

I agree with the others: given the dark square weaknesses in White's game, I would have played 7...Bc5 with hardly a second's thought. E.g. 7...Bc5 8.Na4? -(hoping to force a bishop exchange anyhow)- 8...Bxg1 9.Rxg1 Qh4ch 10.Kd2 Qxh2 11.Be2 Qf4ch with a fine game for Black.

here, white has created holes and structural instability

I would think 9...d6 would be better. The trade of LSB's if White's Bishop were to go to h3 would also show some light square weakness on the kingside as well as force the knight to go to the rim as well as win the f3 pawn.

I prefer 10.Bg2, but this move is OK.

This was always going to be the problem with the move g2-g3, which is why 10,Bg2 was preferable to 10.Bh3. The only way I can see to make it good is 11.Kf2.

Here I would suggest 13.Qd2 and allow the Queens to come off. A pawn down, White will at least have contained the damage to manageable proportions.

IMO, 13.a4 was an unnecessary move, white could move 13.Bg2 straight away. Alternate move was 13.Qd2, because 'safe castling' no longer exists and it makes the white rooks interactive.

Very weakening. This need not be a bad thing if White has a plausible attacking line to follow up.

One can argue opening the long diagonal can activate White's passive bishop on g2...

Exposing the weak h1-a8 diagonal with the discovered attack 19. Rxf7! is the way to go. 19...Kxf7? wins you a piece with 20. Bxd5+. If black simply moves 19...Rd6 to defend the knight, then 20. Rxc7 begins to show the weakness of Black's position. For example, 20...Rc8 21. Rxc8 Bxc8 22. Qxb6 looks bad for Black.

19.Rxf7! (as already noted by rookie879) would throw the game wide open - a fairly desirable outcome from White's point of view, given the situation. As I can't find a decent reply for Black, it would have been the turning point of the game! It just goes to show how loose Black's position had become.